Fused Fixed-Point Arithmetic Unit for Radix-4 DIT FFT Implementation
M B Veena, R A Madhuri, S V Chinmayi, Aishwarya S K · 2022 IEEE 2nd Mysore Sub Section International Conference (MysuruCon) · 2022
Fast Fourier Transform (FFT) is one of the fastest and most efficient algorithms frequently used in DSP applications. This paper presents a design method to compute Radix-4 DIT-FFT for complex fixed-point input using Fused Arithmetic operations. The two fused operations are Fused Add Subtract (FAS) and Fused Dot Product (FDP). In the conventional butterfly computation of FFT, the multiplication, addition and subtraction operations occur in separate blocks and in serial fashion which leads to low throughput, increased area and reduced computational speed. The proposed work is more efficient than the conventional radix-4 butterfly computation in terms of power, area and speed. The design is simulated using Xilinx Vivado and it occupies 46% less area and consumes 2.7% less power.